Morpheme Weight Filtering for High-Speed Email Document Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional document information extracting systems fail to efficiently process document data, particularly in monitoring emails, due to inadequate consideration of increasing processing speed and morpheme diversity.
Innovation Solution
An information processing device and method that utilize a learned model generated by machine learning to determine morpheme weights based on morphological analysis, deleting features with weights below a threshold, and calculating relevance scores for email monitoring, thereby enhancing processing speed and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional document information extracting systems are used to monitor emails, then the system can process document data, but the processing speed is insufficient and cannot handle increasing morpheme diversity efficiently
Solution Approach 1:
The patent segments the document processing task by separating morpheme extraction from full document analysis. The morpheme extraction unit extracts only essential morphemes (words, phrases, or expressions) that are relevant to the analysis target, rather than processing all morphemes in the document. This segmentation enables high-speed processing while maintaining the ability to handle diverse morphemes by focusing computational resources on relevant segments.
Solution Approach 2:
The patent extracts only the necessary morphemes from the document data based on the analysis target, rather than processing all morphemes. The morpheme extraction unit selectively extracts morphemes that contribute to the analysis, deleting or ignoring irrelevant ones. This extraction approach improves processing speed by reducing the volume of data to be processed while maintaining adaptability to handle diverse morpheme types through targeted extraction based on the analysis target.
2Measurement precision
If all morphemes are processed in document data, then comprehensive analysis is achieved, but processing speed decreases and efficiency is reduced
Solution Approach 1:
The patent applies local quality by making the morpheme extraction process target-specific. Different analysis targets (e.g., spam detection, sentiment analysis, topic classification) receive different sets of extracted morphemes tailored to their needs. The morpheme extraction unit adjusts the extraction criteria and focus based on the analysis target, ensuring that each analysis receives the appropriate level and type of morpheme detail without processing all possible morphemes universally.
3Measurement precision
If a learned model determines morpheme weights based on morphological analysis, then relevance scoring accuracy is improved, but the system complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing morpheme weights in the learned model before actual document analysis. The morpheme weights are determined in advance through training on labeled data, so that during runtime, the system only needs to retrieve and apply these pre-computed weights rather than performing complex weight calculations for each document. This preliminary preparation reduces runtime complexity while maintaining high scoring accuracy.
Solution Approach 2:
The patent introduces a learned model as an intermediary between raw morpheme extraction and relevance scoring. This intermediary component (the learned model with pre-determined morpheme weights) simplifies the overall system by providing a ready-made mechanism to evaluate morpheme importance, eliminating the need for complex real-time weight determination algorithms in the main processing path.
Data Source
AI summary
An information processing device includes: a model obtaining unit configured to obtain a learned model generated by machine learning that includes: determining a weight of a morpheme in a model, in accordance with a feature determined using a result of morphological analysis; and deleting, from input data of the model, the feature corresponding to the morpheme having the weight determined to be smaller than, or equal to, a threshold value; an obtaining unit configured to obtain document data ; a feature determining unit configured to determine the feature to be input to the learned model, in accordance with the result of the morphological analysis; an inference processing unit configured to input the feature to the learned model, to calculate a score indicating a degree of relevance between the document data and an event; and a display control unit configured to perform display control using the score.